Automatic marketing strategy generation system based on intelligent service

Abstract

The invention discloses an automatic marketing strategy generation system based on intelligent service, and relates to the technical field of intelligent marketing, and the system comprises a basic template layer, a dynamic adaptation layer, a creative generation layer, and a multi-target strategy selection layer. Through a constructed hierarchical strategy engine of a basic template layer, a dynamic adaptation layer, a creative generation layer and a multi-target strategy selection layer, a single strategy generation mode of a traditional fixed template is broken through, the basic template layer uses an industry marketing knowledge base to provide a standardized strategy generation starting point, the dynamic adaptation layer accesses multi-source heterogeneous data in real time, and the multi-target strategy selection layer is used for providing a standardized strategy generation starting point. According to the method, a creative generation layer is established, strategy parameters are dynamically corrected, market changes can be quickly responded, the creative generation layer automatically generates personalized marketing content by means of a multi-modal generation technology, and a gradient type strategy generation mode is formed from basic framework establishment to real-time parameter adjustment to creative content output through the dynamic layered architecture. And a brand new technical path is provided for marketing strategy generation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent marketing, in particular to an automatic marketing strategy generation system based on intelligent services. BACKGROUND

[0002] In today's competitive business environment, the formulation of marketing strategies plays a crucial role in the survival and development of enterprises. With the rapid development of information technology and the continuous evolution of consumer markets, traditional marketing strategies have become difficult to meet the increasingly complex business needs of enterprises. Enterprises need to dynamically adjust marketing strategies based on real-time market data and consumer behavior to achieve multi-objective balance, such as improving sales, expanding brand awareness, and enhancing user satisfaction, in order to adapt to the rapid changes in the market environment and maintain a competitive advantage. This demand has prompted the urgent need for more intelligent and automated marketing strategy generation systems.

[0003] Currently, the existing related patent technologies have significant shortcomings in strategy generation and evaluation. For example, the classification CN119762145A provides a marketing strategy generation method and system based on big data, which mainly focuses on data collection and simple analysis, but lacks dynamic correction and real-time adaptation functions. It only uses market data and user information to generate fixed strategies, which lacks flexibility in the face of market fluctuations and changes in consumer preferences. The creative content generated is often isolated and not organically integrated with the overall marketing strategy, nor has it undergone rigorous multi-objective optimization evaluation, making it difficult to ensure balanced development of marketing activities in multiple key indicators.

[0004] In summary, the technical bottlenecks of current automatic marketing strategy generation systems are concentrated in the rigid strategy generation mode, the lack of gradient coverage from standardization to personalization, and the inability to meet the diversified marketing needs of enterprises and adapt to differentiated market competition. Therefore, there is an urgent need to build a more adaptive and competitive automatic marketing strategy generation scheme. SUMMARY

[0005] The present application aims to overcome the shortcomings of the prior art and provides an automatic marketing strategy generation system based on intelligent services. It breaks through the single strategy generation mode of traditional fixed templates by constructing a hierarchical strategy engine with a basic template layer, a dynamic adaptation layer, a creative generation layer, and a multi-objective strategy selection layer. The basic template layer uses an industry marketing knowledge base to provide a standardized strategy generation starting point. The dynamic adaptation layer accesses multiple heterogeneous data sources in real time and implements dynamic correction of strategy parameters through differential evolution, enabling rapid response to market changes. The creative generation layer automatically generates personalized marketing content with the help of multi-modal generation technology. This dynamic layered architecture forms a gradient strategy generation mode from basic framework construction to real-time parameter adjustment to creative content output, providing a new technical path and architectural paradigm for marketing strategy generation.

[0006] The application provides the following technical solutions to solve the above technical problems: an automatic marketing strategy generation system based on intelligent services, which comprises a basic template layer, a dynamic adaptation layer, a creative generation layer, and a multi-target strategy selection layer. The basic template layer collects and organizes general strategy frameworks of various marketing activities in the industry, classifies and labels the collected strategy frameworks, establishes a basic template library, and when the system receives a marketing task, the basic template layer matches and extracts the corresponding general strategy framework from the basic template library according to the basic information of the marketing task, and provides a basic framework for strategy generation. The dynamic adaptation layer collects real-time multi-source heterogeneous data, including market data, user data, and enterprise data, analyzes the collected real-time data, mines the change rules in the data, and dynamically corrects the parameters in the general strategy framework provided by the basic template layer according to the data analysis results. The creative generation layer uses natural language processing and image generation technology to generate personalized marketing strategies according to the adjusted strategy parameters of the dynamic adaptation layer. The multi-target strategy selection layer is used to set multiple marketing targets and corresponding weights, input the personalized marketing strategies generated by the creative generation layer into a multi-target optimization algorithm for comprehensive evaluation and sorting, and output the optimal marketing strategy combination.

[0007] Further, the basic template layer comprises a strategy framework collection module, a framework classification and labeling module, and a framework matching and extraction module. The strategy framework collection module is used to collect and organize general strategy frameworks of various marketing activities in the industry, including promotion activity strategies, advertising strategies, member marketing strategies, course sales strategies, and information distribution strategies. The framework classification and labeling module classifies and labels the collected general strategy frameworks, and embeds multi-dimensional configurable strategy parameters in each strategy framework, including discount strength, activity period, and target customer group, and establishes a basic template library. The framework matching and extraction module is used to match and extract the corresponding general strategy framework from the basic template library as a generation benchmark for personalized marketing strategies according to the basic information of the marketing task, including industry type, marketing target, and product characteristics.

[0008] Further, the dynamic adaptation layer comprises a data collection module, a data analysis module, and a parameter correction module. The data collection module is used to collect real-time multi-source heterogeneous data of markets, users, and enterprises, the market data includes industry trends and market size, the user data includes user behavior data, user portraits, and user feedback, and the enterprise data includes product inventory, sales data, and marketing budget. The data analysis module analyzes the collected real-time data and mines the changing patterns in the data; The parameter correction module is used to dynamically correct the parameters in the general strategy framework provided by the basic template layer according to the data analysis results.

[0009] Furthermore, the data analysis module cleans and normalizes the raw data acquired by the data acquisition module and calculates the weights according to the preset values. After normalization, all kinds of data Fusion, to obtain a multi-dimensional vector that comprehensively reflects the market situation and ,in, For the Normalized vector of class data, is the data type weight. Based on the fused data, within the time window T, the trend feature Trend and the fluctuation feature Fluct are extracted to measure the change pattern of the data. , used to calculate the trend of data changes over time, the fluctuation characteristics , used to measure the degree of data dispersion, For the The fusion eigenvalue at time t, 、 is the mean of time and characteristics; The parameter correction module is based on the fusion features output by the data analysis module , Trend and Fluct, modify the configurable strategy parameters of the basic template layer, and the modification is based on the comprehensive indicators of marketing effect To optimize the target, calculate the current strategy parameters right Sensitivity , used to reflect the adjustment The degree of influence on the overall effect is combined with trend and volatility characteristics, and the configurable strategy parameters are adjusted using adaptive step size. ,in, are the updated strategy parameters, Indicates the parameters and , is the total number of strategy parameters, is the learning rate and its value is 0.1-0.3, which controls the amplitude of parameter adjustment. The larger the value, the more obvious the parameter change. is the trend response coefficient. When the data shows an upward trend , when the trend is down , used to make the parameter adjustment direction consistent with the data trend, It is a trend feature. is a fluctuation feature, when , it indicates that the data fluctuation is large, at this time, the is increased, a more aggressive parameter adjustment is made, when , the data fluctuation is small, the is reduced, the parameter is adjusted conservatively, the , C represents the conversion rate, that is, the proportion of users who complete the target behavior of purchase and registration in the total access users, R represents the return on investment, which is the ratio of marketing revenue and cost, and measures the profitability of marketing investment, U represents the user complaint rate, which represents the proportion of users who make complaints in the total users, is the target weight, which is input by the marketing target setting module of the multi-target strategy selection layer, and is used to reflect the emphasis of the enterprise on different targets. The updated strategy parameters are fed back to the data analysis module for analyzing the data trend and fluctuation of the next time window and dynamically optimizing the marketing strategy parameters.

[0010] Further, the creative generation layer includes a text generation module, a multimedia material generation module, and a content fusion optimization module. The text generation module is used to generate personalized marketing copy according to the strategy parameters corrected by the dynamic adaptation layer and the characteristics of the target audience by combining natural language processing technology. The multimedia material generation module uses image generation technology to generate multimedia marketing materials such as images and videos that match the marketing theme and the preferences of the target audience. The content fusion optimization module is used to fuse and optimize the multi-modal content of the generated text, images, and videos to form a complete personalized marketing strategy.

[0011] Further, the text generation module creates targeted text content according to the language habits and interest preferences of different target audiences, and the steps are as follows: Parameter analysis and audience analysis: analyze the strategy parameters corrected by the dynamic adaptation layer, extract key information including product characteristics, marketing targets, and promotion intensity, and at the same time, analyze the characteristics of the target audience such as age, gender, region, consumption habits, and interests by combining the target audience information of the user portrait and performing semantic analysis on the target audience by combining natural language processing technology. Copy template selection and adaptation: select a copy template from a pre-stored copy template library according to the strategy parameters and the characteristics of the target audience, the copy template library contains various types of marketing copy templates such as product promotion copy, promotion activity copy, and brand promotion copy, fill the specific information in the strategy parameters into the copy template, and adapt the expression of the copy to make it conform to the language style of the target audience. Copywriting generation and optimization: Further optimize and generate the adapted copywriting. Based on historical text data, the generated copywriting is optimized in terms of grammatical accuracy, semantic coherence, and marketing appeal. The multimedia material generation module uses the Generative Adversarial Network (GAN) algorithm to generate images and videos. GAN consists of a generator and a discriminator. The generator is responsible for generating realistic images, and the discriminator is responsible for judging the authenticity of the generated images. The steps are as follows: Input condition setting: According to the marketing theme and target audience preferences, set the input conditions for image generation, including product appearance characteristics, scene style, and color preferences; Generator generates images: Based on the input conditions, the generator uses a neural network model to generate preliminary images. The generator generates images that meet the input conditions by learning image data; Discriminator evaluation and feedback: The discriminator evaluates the image generated by the generator to determine its authenticity. If the discriminator deems the image unrealistic, it passes feedback information to the generator. The generator adjusts its generation strategy based on the feedback information and regenerates the image. Iterative optimization: Repeat the generation and evaluation process until the discriminator cannot evaluate the generated image, at which point the generated image is considered a marketing image.

[0012] Furthermore, the multi-objective strategy selection layer includes a marketing goal setting module, an evaluation index construction module, a multi-objective optimization application module, and a strategy ranking selection module; The marketing target setting module is used to preset multiple marketing targets according to the enterprise's marketing strategy and actual needs, and supports dynamic expansion of target types to generate target sets ,in, Indicates the marketing goals, , assigning normalized weights to each target ,satisfy ,in, Characterization target Priority in corporate marketing strategy; The evaluation index construction module is used to construct corresponding strategy evaluation indicators for each preset marketing goal, forming Dimensional evaluation indicator matrix ,in, Indicates the goals The next Segment indicators, , for each segment indicator Assign sub-weights ,satisfy , used to quantify the contribution of the indicator to its target; The multi-objective optimization application module uses a multi-objective optimization algorithm to comprehensively evaluate the generated multiple marketing strategies and calculate the comprehensive fitness score; The strategy ranking and selection module is used to sort the generated multiple marketing strategies according to the evaluation results of the multi-objective optimization algorithm and generate a strategy priority list.

[0013] Furthermore, the multi-objective optimization application module uses a multi-objective optimization algorithm to convert the output of the creative generation layer A personalized marketing strategy is expressed as , each strategy corresponds to a standardized indicator vector ,in Representation Strategy In the Target Normalized scores of indicators and calculation of comprehensive fitness scores , For strategy The comprehensive score of Target The weight of For indicators The sub-weight of .

[0014] Compared with the existing technology, this automatic marketing strategy generation system based on intelligent services has the following beneficial effects: 1. The automatic marketing strategy generation system of the present invention can significantly improve the accuracy and pertinence of marketing strategies. Through the multi-objective strategy selection layer, multiple marketing goals are quantified, evaluated and screened. The system can comprehensively consider multiple factors such as sales growth and user satisfaction improvement, avoiding the limitation of focusing on a single goal in the marketing strategy formulation process. At the same time, combined with the dynamic adaptation layer and the creative generation layer, the system can generate personalized marketing copy, images, and video multimodal content according to the characteristics of the target audience and changes in the market environment, so that marketing information is more in line with the needs and preferences of the target audience. This precise marketing strategy can improve the conversion rate of marketing activities, reduce marketing costs, and bring higher return on investment to enterprises.

[0015] Secondly, the application breaks through the single strategy generation mode of the traditional fixed template by constructing the hierarchical strategy engine of the basic template layer, the dynamic adaptation layer, the creative generation layer and the multi-target strategy selection layer, the basic template layer utilizes the industry marketing knowledge base to provide a standardized strategy generation starting point, the dynamic adaptation layer accesses real-time multi-source heterogeneous data, and realizes dynamic correction of strategy parameters through differential evolution, can quickly respond to market changes, the creative generation layer automatically generates personalized marketing content with the help of multi-modal generation technology, this dynamic hierarchical architecture, from the basic framework construction to the real-time parameter adjustment, to the creative content output, forms a gradient strategy generation mode, provides a new technical path and architecture paradigm for marketing strategy generation.

[0016] Other advantages, objects, and features of the application will be set forth in part in the following specification, and in part will become apparent to those skilled in the art upon examination of the following specification, or can be learned by practice of the application. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0018] Figure 1 It is an operation flow chart of an automatic marketing strategy generation system based on intelligent service; Figure 2 It is a composition schematic diagram of an automatic marketing strategy generation system based on intelligent service. DETAILED DESCRIPTION

[0019] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined application purpose, the following will combine the drawings and the preferred embodiments to specifically describe the specific embodiments, structures, features and effects of the present application.

[0020] Embodiment one: this embodiment takes the marketing promotion of the professional skill training platform for the new occupation course of "digital operation manager" as an example, and details the working principle of the automatic marketing strategy generation system based on intelligent service, such as Figure 2As shown, the composition of the system includes: basic template layer, dynamic adaptation layer, creative generation layer, multi-target strategy selection layer, through the basic template layer to build the industry general strategy framework, the dynamic adaptation layer real-time collection and analysis of training market data, user characteristics and enterprise resources, creative generation layer generates personalized multi-modal marketing content, multi-target strategy selection layer realizes multi-dimensional strategy evaluation and optimization, shows the generation ability of the system from the standardized template to the dynamic precise strategy in the field of vocational education, verifies its effectiveness in improving the course registration rate, brand awareness and user satisfaction, and provides a practical example for intelligent marketing in the vocational skill training industry.

[0021] The basic template layer collects and organizes the general strategy framework of various marketing activities in the industry, classifies and labels the collected strategy framework, and establishes a basic template library. When the system receives a marketing task, the basic template layer matches and extracts the corresponding general strategy framework from the basic template library according to the basic information of the marketing task, providing a basic framework for strategy generation, including strategy framework collection module, framework classification and labeling module, and framework matching and extraction module. Through the vocational skill training platform, the "digital operation teacher" course is launched, covering core skills such as e-commerce operation, data analysis, and new media marketing. The target audience is job seekers and career changers. The strategy framework collection module of the basic template layer collects general strategy frameworks in the field of vocational skill training through industry reports, competitor research, and vocational education interviews, including: course promotion strategy: such as experience class lead, limited price, old lead new course; advertising strategy: vertical recruitment platform advertising, industry community promotion, KOL (industry blogger / lecturer) cooperation; membership marketing strategy: learning points system, VIP exclusive counseling, course combination discount; information distribution strategy: industry trend white paper release, free public class live broadcast, student success case sharing; the framework classification and labeling module classifies the collected strategy framework, divides it into "course traffic acquisition", "conversion rate improvement", "user retention and repurchase" according to marketing goals, and configures adjustable parameters for each framework. In this embodiment, the "experience class lead strategy framework" is taken as an example, the applicable customer group (potential users interested in digital operation and without experience), the activity cycle, the experience class duration, the price, and the supporting services (free data package, tutor counseling) are labeled. After labeling, a basic template library containing strategy frameworks is established, each framework is accompanied by a parameter description document; the framework matching and extraction module filters based on the template of the training task. When the platform inputs the marketing task information: industry type, marketing goal, product characteristics, the framework matching and extraction module filters the core framework from the template library as the generation benchmark of personalized marketing strategy through keyword matching and semantic analysis.

[0022] The dynamic adaptation layer collects multi-source heterogeneous data in real time, including market data, user data and enterprise data, analyzes the collected real-time data, explores the change patterns in the data, and dynamically modifies the parameters in the general strategy framework provided by the basic template layer according to the data analysis results. It includes data collection module, data analysis module and parameter correction module. The data collection module captures three types of data in real time: market data: vocational training industry trends, competitive course pricing, and career change dynamics; user data: target audience behavior data, user portraits, and user feedback; enterprise data: platform teacher reserves, course development progress, and marketing budget. The data is integrated into the data middle platform through API interfaces, enterprise CRM systems, manual entry, etc. The data analysis module cleans and normalizes the original data obtained by the data collection module and calculates the weights according to the preset weights. After normalization, all kinds of data Fusion, to obtain a multi-dimensional vector that comprehensively reflects the market situation and ,in, For the Normalized vector of class data, is the data type weight. Based on the fused data, within the time window T, the trend feature Trend and the fluctuation feature Fluct are extracted to measure the change pattern of the data. , used to calculate the trend of data changes over time, the fluctuation characteristics , used to measure the degree of data dispersion, For the The fusion eigenvalue at time t, 、 is the mean of time and features; the parameter correction module is based on the fusion features output by the data analysis module , Trend and Fluct, modify the configurable strategy parameters of the basic template layer, and the modification is based on the comprehensive indicators of marketing effect To optimize the target, calculate the current strategy parameters right Sensitivity , used to reflect the adjustment The degree of influence on the overall effect is combined with trend and volatility characteristics, and the configurable strategy parameters are adjusted using adaptive step size. ,in, are the updated strategy parameters, Indicates the parameters and , is the total number of strategy parameters, is the learning rate and has a value of 0.1-0.3, is the trend response coefficient. When the data shows an upward trend , the downward trend , is a trend feature, is a volatility feature, the , C represents the conversion rate, i.e. the proportion of users completing the target behavior of purchase, registration to the total access users, R represents the return on investment, which is the ratio of marketing revenue and cost, U represents the user complaint rate, indicating the proportion of users who make complaints in the total users, is the target weight, the updated strategy parameter is fed back to the data analysis module for analyzing the data trend and volatility of the next time window and dynamically optimizing the marketing strategy parameter.

[0023] The creative generation layer adopts natural language processing and image generation technology to generate personalized marketing strategies according to the adjusted strategy parameters of the dynamic adaptation layer. The creative generation layer includes a text generation module, a multimedia material generation module, and a content fusion optimization module. The text generation module is used to generate personalized marketing copy by combining natural language processing technology and the corrected strategy parameters of the dynamic adaptation layer and the characteristics of the target audience. The steps are as follows: Parameter analysis and audience analysis: analyze the corrected strategy parameters of the dynamic adaptation layer, extract key information, including product characteristics, marketing goals, and promotion intensity. At the same time, analyze the characteristics of the target audience, such as age, gender, region, consumption habits, and interests, by combining the target audience information of the user portrait and performing semantic analysis on the target audience using natural language processing technology. Copy template selection and adaptation: select a copy template from the pre-set copy template library according to the strategy parameters and target audience characteristics. The copy template library contains various types of marketing copy templates, such as product promotion copy, promotion activity copy, and brand promotion copy. Fill in the specific information in the strategy parameters into the copy template and adapt the expression of the copy to match the language style of the target audience. Copy generation and optimization: further optimize and generate the adapted copy. Learn from historical text data to optimize the generated copy in terms of grammatical accuracy, semantic coherence, and marketing appeal. The multimedia material generation module uses the generative adversarial network (GAN) algorithm to generate images and videos. GAN consists of a generator and a discriminator. The generator is responsible for generating realistic images, and the discriminator is responsible for judging the authenticity of the generated images. The steps are as follows: Input condition setting: set the input conditions for image generation according to the marketing theme and target audience preferences, including product appearance characteristics, scene style, and color preference. Generator generates images: the generator generates preliminary images using a neural network model based on the input conditions. The generator learns image data to generate images that meet the input conditions. Discriminator evaluation and feedback: The discriminator evaluates the image generated by the generator to determine its authenticity. If the discriminator deems the image unrealistic, it passes feedback information to the generator. The generator adjusts its generation strategy based on the feedback information and regenerates the image. Iterative optimization: Repeat the generation and evaluation process until the discriminator cannot evaluate the generated image, at which point the generated image is considered a marketing image.

[0024] The content fusion and optimization module is used to fuse and optimize the generated multimodal content of text, images, and videos to form a complete personalized marketing strategy for multiple marketing contents.

[0025] The multi-objective strategy selection layer is used to set multiple marketing objectives and corresponding weights, input the personalized marketing strategies generated by the creative generation layer into the multi-objective optimization algorithm for comprehensive evaluation and ranking, and output the optimal marketing strategy combination, including the marketing objective setting module, the evaluation index construction module, the multi-objective optimization application module, and the strategy ranking selection module. The marketing objective setting module is used to preset multiple marketing objectives according to the company's marketing strategy and actual needs, and supports dynamic expansion of objective types to generate objective sets. ,in, Indicates the Marketing goals, , assigning normalized weights to each target ,satisfy ,in, Characterization target Priority in the enterprise marketing strategy; the evaluation indicator construction module is used to construct corresponding strategy evaluation indicators for each preset marketing goal, forming Dimensional evaluation indicator matrix ,in, Indicates the goals The next Segment indicators, , for each segment indicator Assign sub-weights ,satisfy , used to quantify the contribution of the indicator to the target, and use the multi-objective optimization algorithm to convert the output of the creative generation layer A personalized marketing strategy is expressed as , each strategy corresponds to a standardized indicator vector ,in Representation Strategy In the Target Normalized scores of indicators and calculation of comprehensive fitness scores , For strategy a comprehensive score of the target, a weight of the target a sub-weight of the index, a sub-weight of the index, The strategy ranking selection module ranks the generated multiple marketing strategies according to the evaluation results of the multi-objective optimization algorithm, generates a strategy priority list, and takes the comprehensive fitness score as the final selected personalized marketing strategy scheme.

[0026] To sum up, through the marketing practice of vocational skill training courses, the system collects and analyzes market, user and enterprise internal data through a dynamic hierarchical strategy generation mechanism, and combines multi-objective optimization strategy evaluation and selection to generate personalized marketing strategies, thereby improving the accuracy and coverage of marketing strategies.

[0027] Embodiment two: as shown in the embodiment, a process for formulating personalized marketing strategies through an automatic marketing strategy generation system based on intelligent services is provided, and the specific steps of the process are: Figure 1 extracting a general marketing strategy framework from an industry knowledge base to establish a basic template library; classifying the templates according to industry, target, and product characteristics, labeling configurable parameters, and forming a structured template library; According to the input marketing task basic information, retrieve the general strategy framework with the highest matching degree from the template library as the generation benchmark; Real-time collection of market data, user data, and enterprise data; Clean the original data, fuse multi-source data according to the preset weight, and form a feature vector that comprehensively reflects the market situation; Analyze the time trend and fluctuation amplitude of the fused data to identify market change rules and user demand dynamics; According to the market characteristics and marketing effect targets, dynamically adjust the strategy parameters in the template; Combine user portraits and real-time behavior data to analyze the language style and content preferences of the target audience; Based on natural language processing technology, select the adaptive template from the copywriting template library, fill in the strategy parameters, and optimize the grammar, semantics, and appeal; Use image generation and video production technology to generate visual materials according to marketing themes and audience preferences; Integrate copywriting, images, and video content to ensure style consistency and information complementarity, forming a complete personalized marketing content package; Define multiple marketing targets and assign priority weights to each target; Establish a detailed evaluation index for each target and specify the index calculation method and sub-weight; According to the evaluation results of the multi-objective optimization algorithm, the strategy ranking selection module ranks the generated multiple marketing strategies, generates a strategy priority list, and takes the comprehensive fitness score as the final selected personalized marketing strategy scheme. A multi-objective optimization algorithm is used to quantitatively evaluate the generated multiple marketing strategy schemes, and a comprehensive fitness score is calculated; According to the evaluation results, the schemes are ranked, the optimal strategy combination is selected, and an executable personalized marketing strategy is output.

[0028] The above is only a preferred embodiment of the present application, and does not limit the present application in any form. Although the present application has been disclosed as above with a preferred embodiment, it is not intended to limit the present application. Any person skilled in the art can make some changes or modifications to the above disclosed technical content to obtain equivalent embodiments with equivalent changes, without departing from the scope of the technical solutions of the present application. Any modification, equivalent change and modification of the above embodiments made in accordance with the technical essence of the present application, without departing from the technical solution content of the present application, still belongs to the scope of the technical solutions of the present application.

Claims

1. An automatic marketing strategy generation system based on intelligent services, characterized in that: The system consists of: basic template layer, dynamic adaptation layer, creative generation layer, and multi-objective strategy selection layer; The basic template layer collects and organizes the common strategy frameworks of various marketing activities in the industry, classifies and labels the collected strategy frameworks, and establishes a basic template library. When the system receives a marketing task, the basic template layer matches and extracts the corresponding common strategy framework from the basic template library based on the basic information of the marketing task, providing a basic framework for strategy generation; The dynamic adaptation layer collects multi-source heterogeneous data in real time, including market data, user data, and enterprise data, analyzes the collected real-time data, explores the changing patterns in the data, and dynamically modifies the parameters in the general strategy framework provided by the basic template layer based on the data analysis results; The creative generation layer uses natural language processing and image generation technology to generate personalized marketing strategies based on the strategy parameters adjusted by the dynamic adaptation layer; The multi-objective strategy selection layer is used to set multiple marketing objectives and corresponding weights, input the personalized marketing strategies generated by the creative generation layer into the multi-objective optimization algorithm for comprehensive evaluation and ranking, and output the optimal marketing strategy combination.

2. The automatic marketing strategy generation system based on intelligent service according to claim 1 is characterized in that: The basic template layer includes a strategy framework collection module, a framework classification and annotation module, and a framework matching and extraction module; The strategy framework collection module is used to collect and organize the general strategy frameworks of various marketing activities in the industry, including promotional activity strategies, advertising delivery strategies, membership marketing strategies, course sales strategies, and information delivery strategies; The framework classification and annotation module classifies and annotates the collected general policy frameworks, and builds multi-dimensional configurable policy parameters for each policy framework, including discount intensity, activity period, applicable customer groups, and establishes a basic template library; The framework matching and extraction module is used to match and extract the corresponding general strategy framework from the basic template library as a generation benchmark for personalized marketing strategies based on basic information of the marketing task, including industry type, marketing objectives, and product characteristics.

3. The automatic marketing strategy generation system based on intelligent service according to claim 1 is characterized in that: The dynamic adaptation layer includes a data acquisition module, a data analysis module, and a parameter correction module; The data acquisition module is used to collect multi-source heterogeneous data from the market, users, and enterprises in real time. The market data includes industry trends and market size; the user data includes user behavior data, user portraits, and user feedback; and the enterprise data includes product inventory, sales data, and marketing budgets. The data analysis module analyzes the collected real-time data and mines the changing patterns in the data; The parameter correction module is used to dynamically correct the parameters in the general strategy framework provided by the basic template layer according to the data analysis results.

4. The automatic marketing strategy generation system based on intelligent service according to claim 3 is characterized in that: The data analysis module cleans and normalizes the raw data obtained by the data acquisition module and calculates the data according to the preset weights. After normalization, all kinds of data Fusion, to obtain a multi-dimensional vector that comprehensively reflects the market situation and ,in, For the Normalized vector of class data, is the data type weight. Based on the fused data, within the time window T, the trend feature Trend and the fluctuation feature Fluct are extracted to measure the change pattern of the data. , used to calculate the trend of data changes over time, the fluctuation characteristics , used to measure the degree of data dispersion, For the The fusion eigenvalue at the moment, 、 is the mean of time and characteristics; The parameter correction module is based on the fusion features output by the data analysis module , Trend and Fluct, modify the configurable strategy parameters of the basic template layer, and the modification is based on the comprehensive indicators of marketing effect To optimize the target, calculate the current strategy parameters right Sensitivity , used to reflect the adjustment The degree of influence on the overall effect is combined with trend and volatility characteristics, and the configurable strategy parameters are adjusted using adaptive step size. ,in, are the updated strategy parameters, Indicates the parameters and , is the total number of strategy parameters, is the learning rate and has a value of 0.1-0.3, is the trend response coefficient. When the data shows an upward trend , when the trend is down , It is a trend feature. is a fluctuation characteristic, , C represents the conversion rate, which is the proportion of users who complete the target behavior of purchase and registration to the total number of visiting users; R represents the return on investment, which is the ratio of marketing revenue to cost; U represents the user complaint rate, which indicates the proportion of users who file complaints to the total number of users. is the target weight, and the updated strategy parameters The information is fed back to the data analysis module to analyze the data trends and fluctuations in the next time window and dynamically optimize the marketing strategy parameters.

5. The automatic marketing strategy generation system based on intelligent service according to claim 1 is characterized in that: The creative generation layer includes a text generation module, a multimedia material generation module, and a content fusion optimization module; The text generation module is used to combine natural language processing technology to generate personalized marketing copy based on the strategy parameters corrected by the dynamic adaptation layer and the characteristics of the target audience; The multimedia material generation module uses image generation technology to generate multimedia marketing materials such as images and videos that match the marketing theme and the preferences of the target audience; The content fusion and optimization module is used to fuse and optimize the generated multimodal content of text, images, and videos to form a complete personalized marketing strategy for multiple marketing contents.

6. The automatic marketing strategy generation system based on intelligent service according to claim 5 is characterized in that: The text generation module creates targeted text content based on the language habits and interest preferences of different target audiences. The steps are as follows: Parameter parsing and audience analysis: Analyze the strategy parameters modified by the dynamic adaptation layer to extract key information, including product features, marketing goals, and promotional efforts. Combined with the target audience information from the user portrait, analyze the target audience's age, gender, region, consumption habits, and interests. Combined with natural language processing technology, perform semantic analysis of the target audience. Copy template selection and adaptation: Based on strategy parameters and target audience characteristics, a copy template is selected from a pre-set copy template library, which includes various marketing copy templates for product promotion, promotional campaigns, and brand promotion. Specific information from the strategy parameters is then filled into the copy template, and the presentation of the copy is adapted to suit the target audience's language style. Copywriting generation and optimization: Optimize and generate the adapted copywriting. Based on historical text data, the generated copywriting is optimized in terms of grammatical accuracy, semantic coherence, and marketing appeal. The multimedia material generation module uses the Generative Adversarial Network (GAN) algorithm to generate images and videos. GAN consists of a generator and a discriminator. The generator is responsible for generating realistic images, and the discriminator is responsible for judging the authenticity of the generated images. The steps are as follows: Input condition setting: According to the marketing theme and target audience preferences, set the input conditions for image generation, including product appearance characteristics, scene style, and color preferences; Generator generates images: Based on the input conditions, the generator uses a neural network model to generate preliminary images. The generator generates images that meet the input conditions by learning image data; Discriminator evaluation and feedback: The discriminator evaluates the image generated by the generator to determine its authenticity. If the discriminator deems the image unrealistic, it passes feedback information to the generator. The generator adjusts its generation strategy based on the feedback information and regenerates the image. Iterative optimization: The generation and evaluation process is repeated until the discriminator cannot evaluate the generated image, at which point the generated image is considered a marketing image.

7. The automatic marketing strategy generation system based on intelligent service according to claim 1 is characterized in that: The multi-objective strategy selection layer includes a marketing goal setting module, an evaluation index construction module, a multi-objective optimization application module, and a strategy ranking selection module; The marketing target setting module is used to preset multiple marketing targets according to the enterprise's marketing strategy and actual needs, and supports dynamic expansion of target types to generate target sets ,in, Indicates the marketing goals, , assigning normalized weights to each target ,satisfy ,in, Characterization target Priority in corporate marketing strategy; The evaluation index construction module is used to construct corresponding strategy evaluation indicators for each preset marketing goal, forming Dimensional evaluation indicator matrix ,in, Indicates the goals The next Segment indicators, , for each segment indicator Assign sub-weights ,satisfy , used to quantify the contribution of the indicator to its target; The multi-objective optimization application module uses a multi-objective optimization algorithm to comprehensively evaluate the generated multiple marketing strategies and calculate the comprehensive fitness score; The strategy ranking and selection module is used to sort the generated multiple marketing strategies according to the evaluation results of the multi-objective optimization algorithm and generate a strategy priority list.

8. The automatic marketing strategy generation system based on intelligent service according to claim 7 is characterized in that: The multi-objective optimization application module uses a multi-objective optimization algorithm to convert the output of the creative generation layer A personalized marketing strategy is expressed as , each strategy corresponds to a standardized indicator vector ,in Representation Strategy In the Target Normalized scores of indicators and calculation of comprehensive fitness scores , For strategy The comprehensive score of Target The weight of For indicators The sub-weight of .

Images